Pulmonary valve (PV) regurgitation is generally considered to be well tolerated by patients and therefore is clinically not very important. However, over the past decade it has become apparent that this may not be the case. Patients who develop progressive pulmonary regurgitation after operations such as correction of tetralogy of Fallot or complex pulmonary atresia may be prone to persistent pulmonary regurgitation immediately after surgery if a valveless conduit is involved or to progressive pulmonary regurgitation if a valved conduit has been used. Progressively increasing pulmonary regurgitation increases susceptibility to arrhythmias, sudden death and right ventricular (RV) dysfunction (1- 2). There is no consensus on the timing of PV replacement in this clinical situation. It is possible that if the PV is replaced early enough, the RV dilation and dysfunction may be reversible (3- 4). Even so, replacement of the regurgitant PV has not consistently produced recovery of the RV function (5). This may be related to irreversible myocardial damage already present before the valve surgery, aggravated by cardiopulmonary bypass. A non-surgical and therefore less traumatic technique of replacing the regurgitant PV may have important advantages. The purpose of this article is to report our early experience of percutaneous PV implantation in humans.

Eight patients—seven children and one adult—with significant pulmonary regurgitation and/or RV outflow tract obstruction were included in the study protocol. The Ethics Committee (CCPPRB, COCHIN, Paris, France) approved the procedure of percutaneous PV insertion. Fully informed consent was obtained from the parents when the patient was a child and from the patient if the patient was an adult.

The seven children ranged in age between 10 and 17 years, mean of 12.14 ± 2.3 years. One patient was previously reported (6). The details of all the patients are included in (Table le1). Three of the children had previously had surgical repair of pulmonary atresia with ventricular septal defect (VSD), three had previously had repair of tetralogy of Fallot, and the seventh had an absent PV syndrome. The initial palliation included one or more systemic-to-pulmonary artery (PA) shunts in six of eight patients, followed by a complete repair that consisted of closing the VSD and inserting a prosthetic conduit between the RV and the PA. Most of the children underwent reoperation for replacement of the initial conduit with a larger one during infancy.

The adult patient (age 38 years) had undergone repair of tetralogy of Fallot at three years of age, followed by two reoperations to replace the PV 15 and 25 years after the original repair.

All patients were symptomatic, with effort intolerance and breathlessness, and needed conduit replacement because of significant stenosis and/or pulmonary regurgitation. Before the procedure, six patients were in New York Heart Association (NYHA) class II, and two were in NYHA class III with cardiomegaly, moderate-to-severe RV dilation and dysfunction on echocardiography.

In all the patients, cardiac catheterization and hemodynamic evaluation were performed, and angiograms were performed in multiple projections in order to locate the position of the stenosis of the PA or its branches and to define precisely the anatomy of the RV outflow tract.

All the procedures were performed under general anesthesia. Cardiac catheterization was performed via the right femoral vein percutaneously. Heparin (100 U/kg) and antibiotics (cephalosporins) were given intravenously at the beginning of the procedure according to standard protocol. After placement of a right coronary catheter in a distal branch PA, a super-stiff exchange guide wire (Amplatz, Meditech, Watertown, Massachusetts) was positioned in the distal PA. This was followed by a manual inflation of a balloon (Zmed II 18x4, Numed Inc., Nicholville, New York) in the RV outflow tract to determine whether there was sufficient room to implant a PV without creating additional obstruction and to see that the stent could be dilated to its maximum required diameter, especially in those patients who already had some stenosis of the previous surgically implanted valve. Moreover, when distal branch pulmonary arterial stenosis was identified, this was dilated first with an appropriate-sized balloon.

An 18-mm biological valve (Venpro), sutured into a platinum stent (Numed Inc.) as previously described (7), was manually crimped onto a delivery system consisting of a balloon-in-balloon catheter. The valve/stent/balloon assembly was frontloaded in an 18-F or 20-F sheath (Numed Inc.). Thereafter, the stent/balloon assembly was withdrawn fully into a protective sheath on the delivery system. The whole assembly was then passed over the guide wire and advanced into the PA. Once in an acceptable position, the sheath covering the stent was withdrawn to expose the stent crimped on the balloon. The inner balloon was initially inflated followed by the outer balloon, thus deploying the valved stent. The inner and the outer balloons were then rapidly deflated, and the balloon catheter was removed from the patient, leaving the guide wire still in place. Angiographic and hemodynamic evaluation were repeated before the guide wire was finally removed through a right coronary catheter.

After discharge, each patient was seen at regular intervals for clinical evaluation, including physical examination, 12-lead electrocardiography and chest X-ray, and to evaluate the competence of the implanted valve by cross-sectional and color Doppler echocardiography.

Right ventricular angiography before valve/stent implantation showed severe calcification in the previous valved conduit in Patients 1, 3, 5 and 8. In Patients 2 and 4, the stenosis was located at the bifurcation of the PA and at the insertion of the conduit into the RV respectively. Patients 2 and 4 had dilation of branch PA stenosis before the valve/stent implantation. Patient 3 had a severe stenosis of the left PA with multiple pulmonary stenosis distally. In Patients 6 and 7, the stenosis was located in the area of the valve.

The results of the hemodynamic studies are summarized in (Table le2). The valve/stent was successfully implanted in all eight patients. Immediately after implantation, the hemodynamic and angiographic evaluation confirmed competence of the newly implanted valve in six patients (Figure 1). In Patient 4, the stent/valve/delivery system completely obstructed the right outflow tract before the deployment of the valved stent during the placement of the device. This resulted in sudden deterioration with hypotension and severe cyanosis, forcing us to react rapidly. Because we supposed that the valved stent was correctly positioned, we decided to deliver rather than pulling back the device in the RV. Unfortunately, the valve was deployed slightly lower than intended, lying a bit in the infundibulum. In Patients 4 and 5, an insignificant paraprosthetic regurgitation could be visualized because of suboptimal positioning of the valved stent. The relief of the conduit obstruction was partial in Patients 1, 4 and 7.

The fluoroscopy time ranged from 28 to 129 min with a mean at 52 min. The time for the procedure of valve/stent implantation improved significantly after the first two cases. Three patients developed pyrexia after the procedure. The pyrexia lasted for 24 h and resolved spontaneously. No evidence of infection was found in these patients.

All the patients were discharged between one and five days after the procedure with aspirin at low dose. Echocardiography immediately before discharge confirmed competence of the implanted valve in all patients (Figure 2). The slight paraprosthetic leak in Patients 4 and 5 disappeared on color Doppler echocardiography.

At the latest follow-up, ranging from five to 16 months (mean 10.1 months), all the patients had subjective improvement of their symptoms, and color Doppler echocardiography showed a fully competent PV. A reduction of RV size and an improvement of systolic function of the RV were suspected on transthoracic echocardiography evaluation. In the adult patient, there was progressive improvement of the RV function after the implantation, with reduction in the dilation of the RV. Echocardiography also showed that there was persistent elevation of the RV systolic pressure because of partial relief of the conduit obstruction in Patients 1, 3, 4 and 7. The systolic RV pressure tended to be lower echographically during the follow-up compared with the pressure just after the implantation (Table le2). Six patients were in NYHA class I, and the remaining two patients were in class II.

No stent migration occurred. Two stents showed a single fracture of a weld between the wires, which was probably due to difficulties at the time of passing the stent/valve assembly through the skin. There were no clinical sequelae from this.

Prosthetic valved or valveless conduits have been used surgically to establish continuity between the RV and the PA during complete repair of some congenital heart defects, such as pulmonary atresia associated with a VSD or in some patients with the tetralogy of Fallot. Over the past 40 years, these prosthetic conduits have included valveless conduits (8) and conduits with xenograft (9- 14) or homograft valves (15- 16), pericardial valves (17) and mechanical valves (18). Despite major advances in terms of durability, the lifespan of prosthetic conduits is limited. Calcific stenosis of the valve and/or accumulation of intimal peel may cause progressive obstruction of the conduit and valve regurgitation. Therefore, those patients who have received surgically implanted prosthetic conduits are committed to multiple reoperations during their lifespan.

During the past 10 years, aortic and pulmonary homografts and valved conduits constructed from patients’ own pericardium have largely replaced porcine valved conduits (19). This resurgence of aortic or pulmonary allografts, whether cryopreserved or fresh, has not proved to be a panacea. The freedom from reoperation remains only partially satisfactory (20). More recently, fresh autologous pericardial valved conduits have emerged as a possible option. These are more easily available and have demonstrated an improved durability. In a recent long-term follow-up study, freedom from reoperation at five and ten years was 92% and 76%, respectively. Freedom from reoperation at 10 years was 100% for conduits larger than 16 mm (19). Five out of 51 patients (9.8%) required a reoperation for conduit replacement three to eight years after the implantation, and two had a balloon dilation for the relief of a conduit obstruction. It appears, despite conduits’ superior long-term durability, that degeneration is still an important problem, and the perfect valved conduit remains elusive.

Percutaneous stenting of conduits has recently emerged as an additional technique to delay the surgical replacement of the conduit (21- 25). This technique may enlarge the stenotic conduits but, by apposing the valve within the conduit against the wall, creates pulmonary regurgitation that may chronically overload the RV and compromise its long-term function. The time to the occurrence of RV failure as a result of pulmonary regurgitation is variable but may depend on the existence of pulmonary hypertension or distal pulmonary stenosis and the degree of RV hypertrophy.

We have developed a non-surgical technique to treat conduit obstruction by replacing the incompetent PV without compromising valve competence (6). A biological valve was harvested from a bovine jugular vein. A section of this vein containing a tri-leaflet valve was prepared and mounted in a stent for percutaneous PV insertion. Sterilization and cross-linking protocols use glutaraldehyde according to industrial standards. The high quality of this valve has been shown in our experimental work (6). Furthermore, the use of this valve is increasing worldwide in surgery for congenital heart defects in humans. Mid-term and long-term results are still missing, however.

Eight patients were included in our study protocol. All were suffering from obstruction and/or insufficiency of the right pulmonary outflow tract for various reasons. In all the patients, valved stents were successfully implanted, and the patients derived clinical benefit during the early follow-up. The fifth patient, severely symptomatic because of RV dysfunction and severe pulmonary regurgitation, was considered a high-risk candidate for surgery and was thus deemed more suited to percutaneous implantation of the valved stent. Within the first few weeks, there was an improvement of the clinical status. At this stage, the recovery of RV function is uncertain. In surgical experience, the valve replacement is sometimes ineffective, and on other occasions an improvement of symptoms and RV function may occur over a long period of time. The non-recovery of the RV function is generally related to irreversible myocardial damage before the surgery or to myocardial damage of the cardiopulmonary bypass in association with an already diseased ventricle.

In the present study, only valves with a diameter of 18 mm inserted in patients with artificial RV outflow tract were evaluated. Because the size of the whole system is at least 18 Fr, percutaneous insertion in children weighing less than 30 kg is delicate. Eighteen- or 20-mm conduits are usually large enough for adults. Therefore, we hypothesized that insertion of an 18-mm valve in such patients would increase the lifespan of the previously inserted conduits and would last through adulthood. However, two major limitations should be considered. First, the technique we propose did not remove the old PV or dysfunctional conduit before inserting the new valve, as in a conventional surgical procedure. Therefore, by adding a covered stent inside the obstructed conduit, we failed to completely relieve the obstruction in three patients. Second, the question of the durability of the bovine jugular vein valve remains to be answered. However, as far as degeneration is concerned, percutaneous and surgical approaches logically share the same inconvenience. By preserving an adequate long-term RV function, percutaneous valve insertion may diminish the number of surgical interventions or delay the need for surgical conduit replacement. Valvular degeneration is likely to occur in the future because of the biological nature of the valve. In this situation, another valve may be inserted inside the first one, similar to “Russian dolls.” Such a technique would create some obstruction in the conduit and would not be effective in patients without enough room to shelter a second device.

Pulmonary regurgitation is generally considered clinically unimportant. Long-term studies have raised concerns about the function of the RV. Preserving an adequate long-term RV function therefore remains a major challenge. We report the development and application of a non-surgical technique to implant a PV, which may have an important role in the management of conduit obstructions and pulmonary regurgitation. This technique potentially reduces the number of surgical interventions required in patients in whom several operations may have been required to replace malfunctioning conduits.

We thank Allen Tower, Michael Martin, Douglas Villnave, Constantino Quijano and Peter Osypka for their cooperation, advice and technical developments, which contributed to the success of this procedure.